ArangoGraphML is the future of data analytics
Enterprise-ready, graph-powered machine learning as a cloud service
"With ArangoGraphML, now we are able to detect fraud much more easily across all these marketplaces, websites and social media platforms. We would not have been able to do this without Arango's technology."
ArangoDB as a foundation for Graph ML
- Scalable - Designed from ground up to scale enterprise use cases
- Simple Ingestion - Easy integration in existing data infrastructure + connectors to all leading data processing and data ecosystems
- Open Source - Extensibility, Community, especially large community maintained library
- NLP Support - Built-In Text Processing, Search, and Similarity Ranking
Machine learning done right
Traditional machine learning misses connections and relationships between data points, which is where graph machine learn shines. But Graph ML is currently only accessible to large enterprises with dedicated teams of data scientists. ArangoGraphML makes it easier to use graph ML to gain deeper insights from your data.
Node classification as a service
Node classification is at the heart of many machine learning tasks. For example, is a node representing a seller account in an online marketplace classified as fraudulent (selling counterfeit goods) or not? ArangoGraphML makes it easier for data scientists to perform this core part of their job via an intuitive user interface, or an API call from other machine learning tools that they use.
MLOps
To help data science teams focus on high-value tasks, ArangoGraphML includes MLOps to simplify the process of creating machine learning pipelines. Features include model training and management, hyperparameter optimization, metadata, lineage tracking of models and other artifacts, metrics, and dataset management.
GPU support
To uncover machine learning insights faster, ArangoGraphML runs on GPUs (graphics processing units). GPUs are silicon chips that can run computation tasks in parallel and therefore much faster than traditional CPUs, providing increased performance when analyzing large, distributed graphs.
